Module 1: Highlights

Language Probability and Representation

1 Lecture 1.1: Course Orientation and Reproducible GenAI Setup

1.1 Highlights

• Framing Generative AI as a probabilistic system, not a deterministic reasoning engine.
• Course structure, deliverables, and reproducible experimentation standards.
• Establishing a structured GenAI workflow (environment, versioning, logging).
• Introduction to AI disclosure files, documentation practices, and accountability.
• Understanding GenAI as a socio-technical system requiring governance.

1.2 Learning Objectives

By the end of this lecture, students will be able to:

• Explain why GenAI systems operate probabilistically.
• Set up a reproducible environment for GenAI experimentation.
• Describe the importance of AI disclosure and documentation.
• Frame generative models as system components within larger workflows.

---

2 Lecture 1.2: Language Probability and Generative Systems

2.1 Highlights

• Language as a probability distribution over token sequences.
• Prediction as the fundamental mechanism behind generation.
• Conditional likelihood and uncertainty in text modeling.
• The relationship between NLP foundations and modern large language models.
• Introduction to entropy and uncertainty as behavioral drivers in generative systems.

2.2 Learning Objectives

By the end of this lecture, students will be able to:

• Interpret generation as next-token probability prediction.
• Explain conditional probability in language modeling.
• Connect uncertainty to hallucination and variability in outputs.
• Understand why NLP theory underpins modern GenAI systems.